AppwriteUtils Package
The AppwriteUtils package simplifies the process of managing data migrations and schema updates for Appwrite projects. It provides a comprehensive toolset for database setup, data conversion, and schema management, all accessible through a simple command-line interface. This package is designed to be easily integrated into your development workflow, requiring minimal setup to get started.
Getting Started
To use AppwriteUtils, first, install the package via npm:
npm install appwrite-utils
Once installed, you should first run the setup command to generate the config, then you can run migration commands directly using npx
(or bunx
, whatever):
npx appwrite-utils-setup
You may generate an example config using (or look in the examples folder)
npx appwrite-utils-setup --example
npx appwrite-utils-migrate --args
Replace --args
with specific arguments for your migration task. For example, to run migrations in a development environment, you might use:
npx appwrite-utils-migrate --dev
Full Args:
--dev --staging --prod --wipe --wipeDocs --wipeUsers --backup --generate --import
Key Features
For an example, go to the examples
folder here in the repo
- User Import: Quickly import users and their associated data by setting a column as the
usersCollectionName
, it will then map any fields with the target key of name
, email
, prefs
, password
, phone
, userId
, and labels
to create a user object. Any additional info will be mapped to the user like normal. Do not include the users attributes in the attribute map for the collection (the ones mentioned above), they will not be set. The created user and the associated document will have the same $id
. - Simplified Setup: Quickly set up and configure your Appwrite project for migrations with minimal manual intervention.
- Automated Database Migrations: Effortlessly create, update, and manage collections and attributes in your Appwrite database.
- Data Conversion and Import: Convert data from various formats and seamlessly import it into your Appwrite project, adhering to new schema definitions.
- Schema Generation: Automatically generate TypeScript types from your Appwrite schemas, ensuring type safety and consistency across your application.
- Comprehensive Utility Functions: Access a wide range of utility functions for file operations, string manipulation, and more, designed to streamline the migration process.
- String Templating System: Need to update a field with a File ID? There's a post-import function for that (including uploading it). Need to reference an item in your import data's JSON fields to update later? Go for it.
"{$id}"
for instance gets replaced with the created documents ID. If it has one, "{id}"
would be replaced by the JSON items id
field, "{dbId}"
the current dataase, "{createdDoc}"
the created document in this import iteration, and more! - Enhanced Data Management: Handle complex migrations with ease, including managing relationships between collections and converting data for various entities.
This package leverages TypeScript for type safety and is configured to work seamlessly with Appwrite. It's built to support complex migration scenarios, making it an essential tool for developers working with Appwrite projects.
Usage
After installing the package, you can run various migration-related tasks using the command line. Here are some examples of commands you might use, reflecting the capabilities as defined in index.ts
:
-
Initialize a New Migration: Set up your database and prepare it for new migrations. This will also generate schemas but will not import data.
npx appwrite-utils-migrate --init
-
Run Migrations in Production: Apply migrations to your production database.
npx appwrite-utils-migrate --prod
-
Run Migrations in Staging: Apply migrations to your staging database.
npx appwrite-utils-migrate --staging
-
Run Migrations in Development: Apply migrations to your development database.
npx appwrite-utils-migrate --dev
-
Wipe Databases: Wipe your databases. Use with caution.
npx appwrite-utils-migrate --wipe
-
Generate Schemas: Generate TypeScript schemas from your Appwrite database collections.
npx appwrite-utils-migrate --generate
-
Import Data: Import data into your Appwrite project from external sources.
npx appwrite-utils-migrate --import
-
Backup Data: Backup your database data.
npx appwrite-utils-migrate --backup
Each command can be combined with others as needed, except for --init
which runs a specific initialization routine including schema generation but not data import. For example, to run migrations in a development environment and import data, you might use:
npx appwrite-utils-migrate --dev --import
By simplifying the migration process, AppwriteUtils enables developers to focus on building their applications, knowing that their data management and schema updates are handled efficiently.
Helper Functions
Part of this package's goal is utility, to make using Appwrite a bit easier (though it is already quite easy). As such, I've added utility functions, and I'll put them here as I make them
getFileViewUrl
- Takes an endpoint
, projectId
, bucketId
, fileId
, and optionally the current users jwt
to generate a URL to view a file in storagegetFileDownloadUrl
- same as the above but... for downloads
Complete List of Converters, afterImportActions, and Validation Rules
Converters
Converters take a value (in the import data) and convert it, before validating it or processing it
If the converter name has [arr]
or [Arr]
anywhere in it, the converter will be run on the array as a whole (assuming it is one), otherwise if it's an array it will run the converter on each item in the array (or string lol)
anyToString(value: any): string | null
anyToNumber(value: any): number | null
anyToBoolean(value: any): boolean | null
anyToAnyArray(value: any, separator?: string): any[]
anyToStringArray(value: any): string[]
trySplitByDifferentSeparators(value: string): string[]
removeStartEndQuotes(value: string): string
splitByComma(value: string): string[]
splitByPipe(value: string): string[]
splitBySemicolon(value: string): string[]
splitByColon(value: string): string[]
splitBySlash(value: string): string[]
splitByBackslash(value: string): string[]
splitBySpace(value: string): string[]
splitByDot(value: string): string[]
splitByUnderscore(value: string): string[]
splitByHyphen(value: string): string[]
pickFirstElement(value: any[]): any
pickLastElement(value: any[]): any
stringifyObject(object: any): string
parseObject(jsonString: string): any
safeParseDate(input: string | number): DateTime | null
removeInvalidElements(input: any[]): any[]
joinValues(values: any[]): string
(Nothing to join them, just merges em, ignore the name, I am lazy)joinBySpace(values: any[]): string
joinByComma(values: any[]): string
joinByPipe(values: any[]): string
joinBySemicolon(values: any[]): string
joinByColon(values: any[]): string
joinBySlash(values: any[]): string
joinByHyphen(values: any[]): string
convertPhoneStringToUSInternational(value: string): +1 formatted string
validateOrNullEmail
- Added because sometimes you want to just nullify or validate the emails for users and what not before validation
Validation Rules
Validation Rules are run after converters, and are there to make sure invalid data doesn't get added to your database
isNumber(value: any): boolean
isString(value: any): boolean
isBoolean(value: any): boolean
isArray(value: any): boolean
isObject(value: any): boolean
isNull(value: any): boolean
isUndefined(value: any): boolean
isDefined(value: any): boolean
isDate(value: any): boolean
isEmpty(value: any): boolean
isInteger(value: any): boolean
isFloat(value: any): boolean
isArrayLike(value: any): boolean
isArrayLikeObject(value: any): boolean
isFunction(value: any): boolean
isLength(value: any): boolean
isMap(value: any): boolean
isSet(value: any): boolean
isRegExp(value: any): boolean
isSymbol(value: any): boolean
isObjectLike(value: any): boolean
isPlainObject(value: any): boolean
isSafeInteger(value: any): boolean
isTypedArray(value: any): boolean
isEqual(value: any, other: any): boolean
isMatch(object: any, source: any): boolean
has(object: any, path: string): boolean
get(object: any, path: string, defaultValue: any): any
After Import Actions
After Import Actions run after the import and do something with the old data, new data, or something else entirely
-
Provided Fields:
{dbId}
- Current database ID{collId}
- Current collection ID{docId}
- Created document ID{createdDoc}
- Created document objectany_string
- You can use any string or thing as a value too! (like for data){some_template_string}
- The templating system allows you to reference anything in the context of the current
data you're working with. So for instance, if your imported item has {ownerId}
and you use {ownerId}
, it'll reference
that old JSON item data in the import.
-
updateCreatedDocument(dbId: string, collId: string, docId: string, data: any): Promise<any>
-
checkAndUpdateFieldInDocument(dbId: string, collId: string, docId: string, fieldName: string, oldFieldValue: any, newFieldValue: any): Promise<any>
-
setFieldFromOtherCollectionDocument(dbId: string, collIdOrName: string, docId: string, fieldName: string, otherCollIdOrName: string, otherDocId: string, otherFieldName: string): Promise<any>
-
createOrGetBucket(bucketName: string, bucketId?: string, permissions?: string[], fileSecurity?: boolean, enabled?: boolean, maxFileSize?: number, allowedExtensions?: string[], compression?: string, encryption?: boolean, antivirus?: boolean): Promise<any>
-
createFileAndUpdateField(dbId: string, collId: string, docId: string, fieldName: string, bucketId: string, filePath: string, fileName: string): Promise<any>
Roadmap
- Automatic function creation for backups
- Import database schema from Appwrite Server to
appwriteConfig.yaml
(this week) - Promise batching to improve speed
- Deduplication checking
- File based migrations
- Fix custom functions, JS is unable to import TS, so need to find the best way to define those
- Add the ability to import between two appwrite projects
Changelog
-
0.9.986: Added join operators for arrays, join those divorcees!
-
0.9.985: Added the ability to import Users with your data, noted in the #key-features, here's the description:
- User Import: Quickly import users and their associated data by setting a column as the
usersCollectionName
, it will then map any fields with the target key of name
, email
, prefs
, password
, phone
, userId
, and labels
to create a user object. Any additional info will be mapped to the user like normal. Do not include the users attributes in the attribute map for the collection (the ones mentioned above), they will not be set. The created user and the associated document will have the same $id
.
-
0.9.984: Added getFileViewUrl
and getFileDownloadUrl
helper functions to generate the URL's needed with an optional JWT from the current user (if the file needs their permission). Also updated resolveTemplate
inside importDataActions
to better not resolve undefined
in some cases
-
0.9.982: Made the file upload able to deal with arrays of file imports
-
0.9.981: Made documentExists
check for the attribute types in the collection due to running into too many problems with arrays
-
0.9.98: Added [arr]
or [Arr]
arguments to converters. If you use this value anywhere in the converter string, it will
process the value as the array it is (if it is one), otherwise it will run the converter over the mapped array (if it is one)
-
0.9.96: Update to safeParseDate
, it wasn't parsing dates very safely...
-
0.9.95: Fixed a context issue with updating documents in their afterActions using their created data
-
0.9.94: Added a bunch of different splitting functions for more fine-grained control and updated how the trySplitByDifferentSeparators
works to fix the logic. Added converters are above. Also made it so converters and validation actions can take arrays for the item, because why not.
-
0.9.93: Moved relationship resolution to the end
-
0.9.92: forgot I can't use stupid import something from '@/utils'
in esbuild
, stupid, I miss Vite :(
-
0.9.91: Added examples to the example setup with update
importDef type and photos from URL's
-
0.9.90: Rewrote the import process entirely to make it more modular and batched
-
0.9.87 - 0.9.89: Ignore these I was messing up the deployment whooooops
-
0.9.86: I was accidentally including all the code in the NPM package, package size reduced from 2.5 MB -> 53 kB lmao
-
0.9.85: Forgot to make basePath
optional in importDefs
, if it's just an array of objects or somethin you don't need it!
-
0.9.84: jk I use targetKey
a lot, so set it you lazy nerds! (myself included)
-
0.9.83: Made targetKey
optional too, whoops
-
0.9.82: Fixed schema to allow oldKey
or oldKeys
-
0.9.81: Added the ability to match a nested objects properties using [any]
in your template string. So if you have a nested object
"RECORDS": [
{
"someObject": {
"someValue": {
"id": 3,
},
"anotherValue": {
"id": 4,
}
}
}
]
You can resolve the ID's by getting the first one using "oldKey": "someObject.[any].id"
with a converter of converters: ["pickFirstElement"]
or get all of them by using "oldKeys": "someObject.[any].id"
-- this also works in fileData
- 0.9.8: Resolved an issue that would have arisen from not checking for the url in the
fileData
before y'all used it (if y'all is anyone) -- also made sure to fix the type defs in CustomDefinitions - 0.9.7: Added the ability for the
path
field in the fileData
of the importDefs
to be a URL, for laziness! - 0.9.69422: Added
type
and updateMapping
optionally to importDefs
so you can run a second file to update the first one, if needed - 0.9.6942: Added
removeInvalidElements
converter - 0.9.69: Added
oldKeys
to importDefs
so you can concatenate multiple keys to one for an array. Also added five new converter functions, anyToStringArray
, pickFirstElement
, pickLastElement
, stringifyObject
, parseObject
, and a new validator, isDefined
for when you just need to know if something is, well, defined (!undefined, !null, and !empty). I also fixed the exports for the types for the custom definitions, my bad! - 0.9.6: Fixed schema error in enum
- 0.9.5: oops I named it
setup
and migrate
lmao, now it's appwrite-utils-setup
& appwrite-utils-migrate
- 0.9.4: Turns out you gotta import js files in modules, whoops
- 0.9.3: Added
bin
section to package.json and "shebang" to top of main.ts
and setup.ts
to enable npx
- 0.9.2: I forget what I did here
- 0.9.1: Ignore this one
- 0.9.1: Added roadmap 😍
- 0.9.0: Initial refactor into AppwriteUtils package for ease of use